There are some cases in which the speech recognition will not be able to determine if the students pronunciation is correct. Below are some of those instances and the known solutions.

As a general rule of thumb, adding more context will make it easier for the speech recognition to match with the text entered into the playlist. The simplest methods for achieving this are to add articles (the drive), pronouns (his drive), or adjectives (nice drive), or adverbs (drive quickly).


Homonyms that have these same pronunciation but different spellings present a challenge for speech recognition. When added to a playlist as single words, it's impossible for the speech recognition to know which word you intend to use. Homonyms are also one scenario where adding more context may not help.

Examples of homonyms:

accept - take in

except - other than

real - factual

reel - roll

right - correct; not left

write - scribble

ring - encircle

wring - squeeze

road - street

rode - past tense of ride


Create an "OR" function using the "+" that will accept both spellings for he homonym. For example when real + reel are entered into a playlist row both words, when recorded by the speech recognition will lead to a successful pronunciation.

Noisy Rooms / Soft Voices

In classroom settings where a large group of students are practicing their pronunciation at once, background interference can make it more difficult for the speech recognition focus on a single students voice.

This can also be true of students who's voices may naturally be more difficult to hear because they are soft or of a tone that can easily blend with other sounds happenings.

Best Solution:

Headphones - encouraging students to use headphones while they complete speaking exercise is a great solution to this challenge. The microphone will be close to their mouths and it reduces feedback noise (from the computer) within the classroom.

Did this answer your question?